🏳️⚧️🦊 34 / Gaming / Tech / Politics / Advocacy / Disability / Autism
there is no practical technology for long term data preservation. you cannot save a computer file in a way that will last more then a decade. all counterexamples are meaningless; they only exist in a lab. anything you can actually buy will last a few years and then just disintegrate
it is continually astonishing to me that we just keep going through our days, doing nothing about this, while it grows more and more unlikely that any proof that we existed at all will survive us
this shit keeps me up at night constantly as someone very interested in archival
i have the copium-fueled mdiscs that i stick all my data on but i know that if they last a hundred years (i sincerely doubt the 1000 year thing) theyre going to be inaccessible to read in likely half that time because drives won't be manufactured anymore and the OS landscape will be very different. ive been thinking about getting a laser and just engraving a bunch of shit really small (still readable with a magnifying glass) but even then you can't engrave animations or anything interactive (and my hobbies are... animation and gamedev. oops.) the 5d laser quartz storage stuff gets new papers every few months but is still super slow on R/W, is a decade+ from market, and the files will also likely not be readable in 100 years even if it does take off. meanwhile thousands of links go dead every day and ill never find half the stuff i'm seeing online now ever again even in my own lifetime
wtf can we even do about this. how do we make something for storing programs, animations, music, that will not only survive for centuries but also function for centuries? how can we do that when we don't even make things to last 5 years anymore?
at this rate we're NEVER going to be the ancient mystical hyperadvanced people that flew too close to the sun and all died leaving behind strange ruins and dangerous technology. none of our shit will be glowing and vaguely functioning in 50 years much less 5000. cant have shit
is RISC-V at all interesting for archival purposes? Like, 500 years from now, after all copyright and patent concerns are non-issues (for one reason or another), will it be easier to create new hardware that runs RISC-V or ARM?
archival concerns that far along very quickly balloon into abstract scifi fantasyland. like, in 500, 1000, 5000 years, will the issue be that no one remembers specifically how to manufacture/use specific cpu architectures, or will the issue be that no human alive actually knows what a cpu is or how it works?
for my money, forget worrying about one architecture over another. i think they're all going to be forgotten someday. we have to hedge our bets on the tried and true method: rosetta stone the fuck out of it. if humans are alive, they'll have a language, and that language will be descended from one today, so engrave a shitton of rocks with identical text in all the languages we know and put them all over the place. dictionaires and instructions to fix/build modern computers and read the files on them.
then,
make a hard drive that lasts 5000 years ,
like, physically.
its at this point i question if all this fuss is even worth it. like whos gonna go through all this data anyway. whats it matter when 10000 years later someone ELSE has to dig up the remains of THOSE people and figure all this stuff out again. whatever. as long as my shit lasts the years i'm here i don't know that i care too much
what I find extra crazy about the "will no human alive know how a CPU works" contains the quite large probability space "yeah we lost CPU tech" but also the much slimmer "what's a computer, the robots handle all the tech stuff now, let me get back to my AR soaps"
we're going to figure out some bullshit like running all of our computations on the ghosts of ants and we're not even going to know what silicon is anymore
mdisc! that's what i was trying to think of...
constantly transferring data from the previous generation of lto tape to the current gen, only for the lto consortium to update the standard in the night
archisyphus
Archival quality paper, with checksums for OCR on every page. The early 1990s source for PGP will live for centuries. (This is a proven long-term storage technology, but arguably does not meet the bar of "practical")
oh this is a very good thought. can you just laser-print onto that, or is there special hardware needed?
as much as it pains me to say it, laser printing is probably less archival than using giclee ink pigments
if you're going for Extended Lifespan you could inkjet onto transparency for a negative, and then make contact prints on FB paper. or platinotype if you want to go the extra mile
The data density is not great. There have been a few implementations of it; one that I knew about is called Optar, and while trying to find its name again via web-searching, I just found another option called PaperBack. The density of those implementations doesn't exceed 1 megabyte per sheet. There's just only so many bits you can fit into 300 dpi.
That's before data loss to toner wear, paper rot, or other physical factors. (Human text tends to be readable much longer because it endures decay and damage better.)
It can be a solution for very specific things that will fit in the available capacity and that you care enough about to carefully preserve the paper copy of. (One thing I noticed in the search results when I was trying to rediscover Optar's name was that there are apparently some cryptocurrency enthusiasts whose interests have lined up with this, as they particularly want to preserve their wallet/private key/etc.) But it's not going to scale up to larger quantities of data and still be feasible to archive.
when the final computer chip turns to dust then finally we shall be free
When our mom died a few years ago my brother and I spent a bunch of time going through boxes of old photos. Too bad everyone who would recognize anyone in any of these photos is dead, we were thinking! But also, thinking that we have all those old photos. Few more years time everyone will have 100,000 photos of the previous year and zero photos of grandma from back when she was alive. At a high school reunion we were showing a handful of photos from back in the late 80s and these days the kids take more photos per day than we did in three years but in 30 years there will be roughly zero surviving.
that was why they had all thos little tubes. in Startrek
there's not much we can do about it because the choice to use more future-resistant mediums is heavily punished (like most progressive choices in our lives). i have it easy with paint and ceramics, but i think the best digital option is ubiquitous, open source, well-documented storage formats along with manual curation of what's really important in locations we actually own; the profit motive inevitably eats anything stored in the clown.
shit, maybe i should come up with a way to turn jpeg into clay lol. EDIT: wtf of course i can print in stain; that tech already exists
Every so often I go off about mircofilm for this reason, but that era is dead and gone
it would be very funny if M-DISC is the only thing that lasts and all the data in the world rots anyway because everyone forgot about M-DISC
anyway i'm glad i take photos on film and have a printer that makes archival-stable prints. my prints are guaranteed to outlast my files, possibly by an order of magnitude
